What I Built
TerraQuest AI is a personalized outdoor-quest generator that turns a vague intention like “I should spend less time on my screen” into a small, specific activity you can start in minutes.
Instead of handing you a generic list of things to do, TerraQuest asks a few quick questions and shapes a quest around your day:
Activity type: for example, a mindful walk
Available time: 5, 10, 15, 30 or 60 minutes
Energy level: from easy and up
Accessibility needs: seated options, low-energy days, flat routes
Interests: what you’d like to notice, such as birds, moss or cloud shapes
Eco-mission: an optional small act of care for nature, or a surprise
Every quest comes with an accessible alternative, so weather, energy or terrain never leaves you without an option.
How it gets people off the screen
Discover: a daily “Today’s invitation” lowers the barrier to starting.
My quests: saved quests you can return to and complete.
Nature journal: a place to record observations and moments from outside.
Passport: a local record of minutes outdoors, quests completed, activities explored and weekly activity, with a progress bar toward a 30-minute goal. Quiet milestone badges such as First Quest, Curious Observer and Seven Outdoor Breaks mark your progress without streak pressure.
Who it’s for: students, remote workers and anyone who spends long hours on a screen and wants a gentle, low-pressure nudge outside. It’s built for all energy levels and abilities, which is why accessibility is an input to every quest rather than an afterthought.
Privacy by design: no account, no precise location and no tracking. Your progress is stored locally on your device.
Demo
🎥 Video walkthrough:
🔗 Live app: https://terraquest-ai-iota.vercel.app
Code
TerraQuest AI
Less scrolling. More exploring.
TerraQuest AI is a local-first nature companion for making small outdoor adventures feel easy to start. This first implementation focuses on the core experience: discovering activities, generating a practical quest, saving it in the browser, completing it in a low-distraction mode, and keeping a nature journal.
Current core features
- Responsive landing / Discover page with activity suggestions and real local impact counters.
- Quest creator with activity, duration, difficulty, and observation-interest inputs.
- Quest Council boundary: Gemma/Ollama generation, structured output validation, and clearly labeled curated fallback.
- Saved, in-progress, completed, searchable-by-tab quest views.
- Active quest mode with checklist and optional timer.
- Nature journal with add, edit, and delete flows.
- Browser
localStoragepersistence for quests and journal entries. - Nature Passport, Eco-Missions, accessibility preferences, photo attachments, export/reset controls.
- Express local API at
POST /api/quests/generateandGET /api/health. - Online/offline status indicator and truthful local-storage messaging.
- Reduced-motion-friendly transitions and keyboard-focus…
How I Built It
TerraQuest AI uses Gemma, an open-weight model, served through Ollama to generate quests from the user’s preferences. The model runs locally on my Mac, so I could build and experiment with open-weight AI without depending on a closed, paid AI API for inference.
How the AI fits in
The quest form collects the user’s activity type, eco-mission, available time, energy level, accessibility needs and interests.
These inputs are combined into a structured prompt and sent to Gemma through Ollama.
Gemma generates a quest: a title, a short invitation, simple steps, something to notice, and an accessible alternative.
Completed quests, minutes outdoors and journal entries are stored locally and feed the Passport and its badges.
Design decisions
Small and specific: a good quest is something you can start right now, not a plan you’ll put off.
Calm, nature-inspired design: the interface should feel like an invitation, not a productivity tool.
Gentle progress: quiet milestones instead of guilt-inducing streaks.
Accessibility built in: low-energy and seated alternatives are part of quest generation from the start.
Local-first: user data stays on the device.
Why Does Open Innovation Matter?
TerraQuest is built on the belief that getting outside should be accessible, private and low-pressure. Open innovation made that possible in ways a closed API wouldn’t have:
Privacy: with a locally run model, a user’s preferences, energy levels and accessibility needs never have to leave their machine.
No paywall or API dependency: I could build, test and iterate freely, without usage costs, rate limits or deprecations from a single vendor.
Control and customization: open weights let me inspect, prompt and tune the model so quests respect real-world constraints like limited time, low energy or mobility needs.
Community potential: anyone can fork the project, swap in a different model, or adapt it for schools, parks and community groups.
A closed API could have generated quests too, but it couldn’t have offered the same privacy, independence and room for others to build on it.
Top comments (0)